Drugs 44 (Suppl. I): 67-73, 1992 0012-6667/92/0 I 00-0067/$3.50/0 © Adis International Limited. All rights reserved. DRSUP3416
Risk-Benefit Aspects of Antihypertensive Drugs H. Holzgreve and M. Middeke University of Munich, Medizinische Poliklinik, Munich, Germany
Summary
Several long term trials using traditional antihypertensive therapy with diuretics and J3-blockers have shown that antihypertensive therapy reduces the overall risk of cardiovascular complications. However, even after several years of therapy the cardiovascular risk in hypertensive patients cannot be lowered to that in the normotensive population. Antihypertensive therapy can reduce the incidence of cerebrovascular complications in patients with hypertension by about 65%. However, the effect of such therapy in preventing coronary events has been disappointing, as these events are 3 to 4 times more common than cerebrovascular complications in hypertensive patients. It is now apparent that adverse pharmacological effects of diuretics and J3-blockers on lipid metabolism persist for many years. Thus, treatment with these agents constitutes a new risk factor for coronary heart disease and may, at least in part, explain the failure of traditional antihypertensive therapy to reduce the incidence of myocardial infarction and sudden death as effectively as that of cerebrovascular accidents. On the other hand, titration of these antihypertensive agents to the lowest possible dose in order to avoid metabolic alterations and subjective adverse effects has frequently resulted in the administration of subtherapeutic doses, particularly for hydrochlorothiazide. Until comparative long term clinical trials with older and newer antihypertensive agents and morbidity and mortality as end-points are completed, the debate on firstline drugs for antihypertensive treatment will not be satisfactorily resolved.
During the past decade, several new drugs have been introduced into antihypertensive therapy, thus providing alternative options for treatment. The multiplicity of substances now available for therapy is naturally a fertile field for debate among those involved in the treatment of hypertension. However, even the recommendations of expert bodies such as the Joint National Committee in the US and the German Hypertension League have undergone a fundamental change during the last 2 decades. During the 1970s and until the mid-1980s, diuretics and /3-blockers were recommended worldwide as first-line drugs for the treatment of hypertension. The above-mentioned organisations currently recommend 4 groups of drugs as being of
equal value: diuretics, /3-blockers, calcium antagonists and angiotensin converting enzyme (ACE) inhibitors (Deutsche Liga zur Bekampfung des hohen Blutdrucks e.V. 1990; Report of the Joint National Committee 1988). One of these classes of agents, the calcium antagonists, was developed in Germany, but for some time these drugs were not commonly accepted there for the treatment of hypertension. This has changed dramatically during recent years (fig. 1). The standard units (millions per year) of /3-blockers and diuretics sold in Germany remained fairly static between 1984 and 1990, but those of ACE inhibitors and calcium antagonists steadily increased. In 1990, the numbers of units of calcium antagonists sold
Drugs 44 (Suppl. 1) 1992
68
2. Lipid Metabolism in Long Term Trials
• j3-Blockers o Diuretics • Calcium antagonists o ACE inhibitors
800
1985
1986
1987
1988
1989
1990
Year Fig. 1. Standard units (in millions) of antihypertensive agents sold per year in Germany between 1984 and 1990. ACE = angiotensin converting enzyme.
were higher than those of ACE inhibitors, diuretics or /3-blockers. Similar trends have been observed in many other countries.
1. Hypertension Intervention Trials There has been much debate in favour of one or another individual drug or substance group_ At times, however, no adequate distinction is made between important, less important, appropriate, inappropriate, or theoretically correct but clinically irrelevant criteria. The selection of antihypertensive agents should be based on evidence that the drug or drugs chosen not only lower blood pressure but also reduce the likelihood of the typical complications of hypertension and thus improve the patient's prognosis. An analysis of the results of the various studies performed to date on the long term efficacy of antihypertensive agents shows that 5 years' antihypertensive drug treatment reduces overall mortality by 33%, cardiovascular mortality by 41%, and the incidence of fatal and nonfatal cerebrovascular events by 51 %. However, the complications of coronary heart disease (i.e. fatal and nonfatal myocardial infarction and sudden cardiac death) are reduced by only 15% and such events occur 3 to 4 times more frequently than cerebrovascular events in patients with hypertension (refer to studies cited in table I).
Overall, data from long term trials (see table I) attest to the outstanding benefit of antihypertensive therapy. However, the results are disappointing as far as coronary heart disease events are concerned, particularly since cardiac complications occur 3 to 4 times as often as cerebrovascular events. Diuretics and /3-blockers were the antihypertensive agents used in these trials and many short term studies (up to several months' duration) have revealed adverse effects of both these drugs on serum lipoproteins (Middeke et al. 1987). The failure of these agents to reduce the incidence of myocardial infarction and sudden death as effectively as the occurrence of cerebrovascular accidents has stimulated interest in lipid metabolism during antihypertensive therapy. Deterioration, normalisation and even improvement in lipoprotein levels have all been observed in a few long term studies of diuretics or /3-blockers over several years (Wilhelmsen et al. 1987; Williams et al. 1983). Possible explanations are that the pharmacological effects on lipoproteins disappear during longer term use, or that other factors influencing serum lipoprotein concentrations, for instance changes in diet, bodyweight or aging, may override any pharmacological effect. Table I. Reduction in mortality and morbidity in major hypertension intervention trials calculated as the percentage reduction in treated groups (extrapolated over 5 years) compared with control or placebo groups [Amery et al. 1985 (EWPHE); Dahlof et al. 1991 (STOP); Hypertension Detection and Follow-up Program Cooperative Group (HDFP) 1979a.b. 1982a.b. 1984a.b. 1985; Management Committee 1980; McFate Smith 1977 (USPH); Medical Research Council Working Party (MRC) 1985. 1992; SHEP Cooperative Research Group 1991; Veterans Administration Cooperative Study Group on Antihypertensive Agents 1967. 1970.1972] Percentage reduction Overall mortality Cardiovascular mortality Cerebrovascular events Coronary heart disease events
33 41 51 15
69
Risk-Benefit Aspects of Antihypertensive Drugs
Patients in the Munich centre who participated in the HAPPHY trial (Wilhelmsen et al. 1987) one of the major randomised long term intervention trials comparing diuretics and {j-blockers provided a unique opportunity to observe the effects of long term treatment and subsequent discontinuation of antihypertensive treatment (Middeke et al. 1990). After 5.2 years of therapy, the mean doses of hydrochlorothiazide and atenolol were 51 and 75mg daily, respectively (table II). Blood pressure was 145/91mm Hg in patients treated with hydrochlorothiazide and 142/ 93mm Hg in patients treated with atenolol; during the 4.6-week post-treatment washout period, blood pressure increased significantly to 165/104 and 159/ 106mm Hg, respectively. After hydrochlorothiazide was discontinued, total cholesterol decreased from 6.4 to 6.0 mmol/L (247 to 231 mg/dl) and low density lipoprotein (LDL) cholesterol from 4.3 to 3.9 mmol/L (167 to 150 mg/dl) [both p = 0.01 vs during treatment], and high density lipoprotein (HDL) cholesterol remained unchanged. Cessation of treatment with the ,a-blocker led to a decrease in LDL cholesterol from 4.2 to 3.9 mmol/L (162 to 150 mg/dl) and an increase in HDL cholesterol from 0.96 to 1.17 mmol/L [37 to 45 mg/dl], both of which were significant changes. These results clearly demonstrate that the lipid changes were reversible after medication was discontinued. On the other hand, the adverse effects of the diuretic and the ,a-blocker had persisted for more than 5 years. Thus, treatment with these agents has introduced a risk factor for coronary heart disease. This increased risk of cardiovascular events may well have reduced the benefit of treatment.
3. Comparison 0/ a Diuretic and a Calcium Antagonist A consequence of these results has been recommendations of dosage reductions, as many side effects of drugs are dose dependent. Whereas 10 to 15 years ago clinicians used daily doses of up to 200mg of hydrochlorothiazide, doses of only 25 or at most 50mg daily are currently recommended. Since the dose-response relationship was claimed
Table II. Blood pressure and serum lipoproteins after long term therapy (5.2 years) with the diuretic hydrochlorothiazide (HCT) or the /3-blocker atenolol. and 4.6 weeks after discontinuation of treatment (Middeke et al. 1990) Diuretic (HCT 51 mg/day)
/3-Blocker (atenolol 75 mg/day)
Number of patients
23
17
Blood pressure (mm Hg) Long term therapy Discontinuation
145/91 165/104*
142/93 159/106*
Total cholesterol (mmol/L) [mg/dl] Long term therapy Discontinuation
6.4 [247] 6.0 [231]**
6.2 [240] 6.0 [233]
LDL cholesterol (mmoI/L) [m9ldl] Long term therapy Discontinuation
4.3 [167] 3.9 [150]**
4.2 [162] 3.9 [150]**
HDL cholesterol (mmol/L) [ma/dl] Long term therapy Discontinuation
1.19 [46] 1.17 [45]
0.96 [37] 1.17 [45]·**
= high density lipoprotein; LDL = low density lipoprotein. * p = 0.001 vs end of therapy; *. p = 0.01 vs end of therapy; ••* p = 0.05 vs end of therapy.
Abbreviations: HDL
to be flat, especially for diuretics, reducing the dosage was not expected to reduce the antihypertensive effect. However, efficacy is dose dependent and it may well be that dose reduction has resulted in administration of less effective doses. For this reason, a study was initiated to evaluate a thiazide diuretic and a calcium antagonist. The I-year comparative therapeutic trial in 369 patients with mild to moderate hypertension (Holzgreve et al. 1989) had a multicentre randomised double-blind design with a single-blind placebo baseline period of 2 weeks. Patients who had been previously treated with antihypertensive agents had an additional 2-week washout period. Hydrochlorothiazide, which has been used in many intervention trials and which is still taken as the gold standard in antihypertensive therapy, was the selected diuretic, and the comparative drug, verapamil, was selected as an example of agents so far
70
Drugs 44 (Suppl. 1) 1992
not used in long term intervention trials in hypertension. Dosage levels of hydrochlorothiazide used were 12.5mg, 25mg and 50mg daily, and verapamil was given at a dosage of 120,240 and 480mg daily. The target blood pressure was defined as a diastolic blood pressure of less than 90mm Hg. If this was not achieved after 2 weeks the next regimen was prescribed. Where even the highest of the 3 dose steps did not result in achievement of target blood pressure, the 2 drugs were combined in 2 dose steps, i.e. initially hydrochlorothiazide 25mg plus verapamil 240mg once daily increasing to 50mg plus 480mg, respectively, on dose step V if required. According to the protocol, patients with mild, moderate and severe hypertension were included. Of the 369 patients who entered the trial, 187 were randomised to treatment with hydrochlorothiazide and 182 to the verapamil group. The randomisation resulted in good comparability between the 2 treatment groups, and both groups had similar systolic and diastolic blood pressures. Mean systolic blood pressures were 152.6 and 152.3mm Hg, in the hydrochlorothiazide and verapamil groups, respectively, and mean diastolic blood pressures were 104.4 and 104.2mm Hg, respectively.
4. Response Rates 4.1 Monotherapy At 8 weeks, the percentage of patients at the target blood pressure, i.e. diastolic less than 90mm Hg, was 42.7% for hydrochlorothiazide-treated patients and 57.7% for verapamil recipients (p < 0.005; fig. 2). At 24 weeks, the values were 26.4% and 46.8% (p < 0.0001) and after 48 weeks 24.7% and 44.8% (p < 0.0001), respectively. These differences show that verapamil significantly was more effective than hydrochlorothiazide at all time points during the study. Furthermore, verapamil was more effective at all dose steps used. Almost twice as many patients on monotherapy with verapamil reached the end of the study compared with monotherapy with hydrochlorothiazide. The fact that more patients remained for a longer time on monotherapy with verapamil is important to consider when analysing the adverse effects of the 2 drugs. In or-
der to determine whether individual patients responded better to one or the other of the medications, subgroup analyses were performed, and comparative results of patients with a blood pressure at entry below and above the median of 103mm Hg, patients aged below 45, 45 to 55 and above 55 years of age, and patients from clinical outpatient departments and from private practices are also shown in figure 2. In all subgroups, patients treated with verapamil had a higher response rate than patients treated with hydrochlorothiazide. Thus, verapamil was more effective than hydrochlorothiazide in all subgroups. 4.2 Combination Therapy It has been suggested that combinations of calcium antagonists and diuretics (or related compounds) are of little value in the treatment of hypertension. Indeed, in several studies there was no significant blood pressure-lowering effect from adding diuretics such as bendroflumethiazide (Cappuccio et al. 1986, 1987, 1991) mefruside (Rosenthal 1983) or chlorthalidone (Salvetti et al. 1991) to nifedipine or amlodipine. In one study only, the addition of hydrochlorothiazide to verapamil caused a further significant fall in blood pressure (MacGregor et al. 1988). Furthermore, it has been postulated that there is a sequential effect, i.e. the antihypertensive effect is improved when a calcium antagonist is added to a diuretic, but not when a diuretic is added to a calcium antagonist (Hansson 1991). As already shown in our study, 24.7% of patients achieved target blood pressure after 48 weeks' monotherapy with hydrochlorothiazide, whereas the success rate after verapamil monotherapy was 44.8% (fig. 3), i.e., when the drugs were used as monotherapy, verapamil was effective almost twice as often as hydrochlorothiazide. Patients who failed to respond to monotherapy were given verapamil and hydrochlorothiazide in combination. The addition of verapamil to patients who were initially randomised to hydrochlorothiazide and did not achieve target blood pressure increased the response in 34.3% of patients. Conversely, addition
Risk-Benefit Aspects of Antihypertensive Drugs
71
70 60 iii'
E 50
.~
iii
"-
'0 40
t
Q> V>
c
30
0
"V> 20 Q>
a:
10 0
All
patients
103 > 103 BP (mm Hg)
~
45-55 > 55 Years 01 age
< 45
In In private clinics practice
Fig. 2. Percentages of patients (all patients and various subgroups) achieving target blood pressure after 8 weeks' monotherapy with hydrochlorothiazide (0) or verapamil (.).
of hydrochlorothiazide to patients who were initially randomised to verapamil and did not achieve target blood pressure increased the response rate in only 17.8% of patients; thus, compared with hydrochlorothiazide, verapamil was not only effective twice as often when used as monotherapy but also when used as a drug for combination. These results clearly show that there is an additive effect from combining verapamil and hydrochlorothiazide. Thus, according to the previous studies there might be a difference between calcium antagonists SUQb as verapamil and the dihydropyridines. In cSu~ study, the antihypertensive efficacy of hydrochlorothiazide was lower than that of verapamil, and more patients achieved target blood pressure after the addition of verapamil to hydrochlorothiazide compared with adding the diretic to the calcium antagonist. It is possible that indings from previous studies that failed to dem.nstrate a statistically significant effect of diuretics n patients already taking a calcium antagonist night have been influenced by the small populaion studied (Cappuccio et al. 1986, 1987, 1991), lr may have resulted from the use of subtherapeuic diuretic doses (Salvetti et al. 1991). As far as a possible sequential effect is concerned, it was to be expected that the total percentage of patients who responded either to monotherapy or combination
therapy would be the same in both groups, assuming an additive effect and excluding any sequential effect. Indeed, the figure for total response rate was almost identical, i.e. 59.0% in the hydrochlorothiazide group and 62.6% in the verapamil group (fig. 3). This result excludes any sequential effect.
5. Adverse Effects and Withdrawals The higher efficacy of verapamil might be due to the fact that in recent years titration to the lowest possible dose of diuretics has been suggested in order to avoid side effects and adverse metabolic alterations, such as those on blood lipoprotein leviii' 75 C
59.0%
0; "- 50
Plus
62.6%
.~
'0 ~
VER
17.8% 34.3% Mono
Q> V>
c 25 0 a.
VER
44.8%
V> Q>
a:
0
HCT group
Verapamil group
Fig. 3. Percentages of patients achieving target blood pressure in groups randomised to 48 weeks' therapy with hydrochlorothiazide (HCT) or verapamil (VER) as monotherapy or the 2 drugs combined.
Drugs 44 (Suppl. 1) 1992
72
els. In the current study, this strategy may well have resulted in administration of doses that cause fewer adverse effects but that reduce the elevated blood pressure less effectively. On the other hand, the doses ofverapamil selected in our study might well be more potent in reducing blood pressure, but at the same time could cause more adverse effects than hydrochlorothiazide. The answers provided by the patients in this double-blind study using a structured questionnaire show that the patients treated with hydrochlorothiazide complained relatively more commonly of urinary frequency and those treated with verapamil reported constipation more frequently, although the differences were not analysed for statistical significance. Potassium salts and potassium-sparing diuretics were prescribed more often for patients in the hydrochlorothiazide group than for patients in the verapamil group. The only significant change in laboratory parameters was a mean fall of 0.311 mmol/L in serum potassium. However, contrary to findings from our earlier studies, in the group of 178 patients receiving hydrochlorothiazide no significant changes in total cholesterol, LDL cholesterol, HDL cholesterol and triglycerides could be observed at 8 weeks. It is suggested that this was due to the lower mean daily dose of hydrochlorothiazide used (32mg) compared with a dose of 51 mg used in our studies cited above (Middeke et al. 1990). The total number of patients withdrawing from the study are shown in figure 4. Administrative reasons such as noncompliance or protocol violations accounted for 30 to 65% of withdrawals across the 4 groups. Increased diastolic blood pressure (> 120mm Hg) was the most frequent adverse effect causing withdrawal among all remaining patients. Similar number of patients were withdrawn from each monotherapy group. However, it must be remembered that because of the greater efficacy of verapamil, more patients remained on verapamil for a longer time than on hydrochlorothiazide and therefore were at a higher risk of adverse effects associated with verapamil. There were 10 withdrawals by patients for whom verapamil had been added to hydrochlorothiazide and 17 dropouts in
"'c:.
25
~ a.
20
ci
15
"5
.s $
~
iii ~
~
25
23
17
10 10
5
"0
-s ~
0
HCT
VER
Monotherapy
HCT
+ VER
VER + HCr
Combination therapy
Fig. 4. Total number of patients withdrawing from randomised therapy during the trial, on monotherapy with hydrochlorothiazide or verapamil or after adding verapamil to hydrochlorothiazide (HCT + VER) or hydrochlorothiazide to verapamil (VER + HCT).
patients who had hydrochlorothiazide added to verapamil therapy. Thus, the withdrawal rate during treatment with verapamil seems no higher than that during treatment with hydrochlorothiazide.
6. Conclusions and Therapeutic Implications In conclusion, findings from this study have shown that smaller doses of the diuretic reduce the incidence of side effects, particularly of adverse lipid alterations, and that the incidence of treatmentrelated adverse effects and withdrawals was essentially the same with the 2 drugs. However, verapamil had a significantly more pronounced antihypertensive effect than hydrochlorothiazide. Physicians should base their therapeutic decisions on currently available data, although even experts may differ in their interpretation. As mentioned above, expert bodies such .as the German Hypertension League and the Joint National Committee currently recommend 4 groups of drugs as first-line agents: diuretics, {j-blockers, calcium antagonists and ACE inhibitors. However, the results presented here must call these recommendations into question - at least as far as diuretics and calcium antagonists are concerned - since they showed that with regard to antihypertensive efficacy and tolerability standard doses of the calcium antagonist were clearly superior to those of the diuretic.
Risk-Benefit Aspects of Antihypertensive Drugs
References Amery A, Birkenhager W, Brixko p, Bulpitt C, Clement D, et al. Mortality and morbidity results from the European Working Party on High Blood Pressure in the Elderly trial. Lancet I: 1349-1354, 1985 Cappuccio FP, Markandu ND, Singer DRJ, Buckley MG, Miller MA, et al. A double-blind crossover study of the e/Tect of concomitant diuretic therapy in hypertensive patients treated with amlodipine. American Journal of Hypertension 4: 297-302, 1991 Cappuccio FP, Markandu ND, Tucker F, Sagnella GA, MacGregor GA. Does a diuretic cause a further fall in blood pressure in hypertensive patients already on nifedipine? Journal of Clinical Hypertension 4: 346-353, 1986 Cappuccio FP, Markandu ND, Tucker FA, Shore AC, MacGregor GA. A double-blind study of the blood pressure-lowering e/Tect of a thiazide diuretic in hypertensive patients already on nifedipine and a beta-blocker. Journal of Hypertension 5: 733-738, 1987 Dahlbf B, Lindholm LH, Hansson L, Scherstein B, Ekbom T, et al. Morbidity and mortality in the Swedish trial in old patients with hypertension (STOP - Hypertension). Lancet 338: 1281-1285,1991 Deutsche Liga zur Bekiimpfung des hohen Blutdrucks e.V. Empfehlungen zur Hochdruckbehandlung in der Praxis und zur Behandlung hypertensiver Notfalle, 9th Ed., 1990 Hansson L. Position of calcium antagonists in antihypertensive combination therapy. Abstract. 2nd International Symposium on Calcium Antagonists in Cardiovascular Care, Basle, 1991 Holzgreve H, Distler A, Michaelis J, Philipp T, Wellek S. Verapamil versus hydrochlorothiazide in the treatment of hypertension: results of a long term double blind comparative trial. British Medical Journal 299: 881-886, 1989 Hypertension Detection and Follow-up Program Cooperative Group. Five-year findings of the Hypertension Detection and Follow-up Program. I. Reduction in mortality of persons with high blood pressure, including mild hypertension. Journal of the American Medical Association 242: 2562-2571, 1979a Hypertension Detection and Follow-up Program Cooperative Group. Five-year findings of the Hypertension Detection and Follow-up Program. II. Mortality by race, sex and age. Journal of the American Medical Association 242: 2572-2577, 1979b Hypertension Detection and Follow-up Program Cooperative Group. Five-year findings of the Hypertension Detection and Follow-up program. III. Reduction in stroke incidence among persons with high blood pressure. Journal of the American Medical Association 247: 633-638, 1982a Hypertension Detection and Follow-up Program Cooperative Group. The e/Tect of treatment on mortality in 'mild' hypertension. New England Journal of Medicine 307: 976-980, 1982b Hypertension Detection and Follow-up Program Cooperative Research Group. The e/Tect of antihypertensive drug treatment on mortality in the presence of resting electrocardiographic abnormalities at baseline: the HDFP experience. Circulation 79: 996-1003, 1984a Hypertension Detection and Follow-up Program Cooperative Group. E/Tect of stepped care treatment on the incidence of myocardial infarction and angina pectoris. 5-year findings. Hypertension 6 (Suppl. I): 198-206, 1984b Hypertension Detection and Follow-up Program Cooperative Group. Five-year findings of the Hypertension Detection and Follow-up Program. Prevention and reversal of left ventricular hypertrophy with antihypertensive drug therapy. Hypertension 7: 105-112, 1985 MacGregor GA, Markandu ND, Singer DRJ. Diuretics and calcium antagonists in hypertension. British Journal of Clinical Practice 60: 58-59, 1988
73
Management Committee. The Australian Therapeutic Trial in Mild Hypertension. Lancet I: 1261-1267, 1980 McFate-Smith W. Treatment of mild hypertension. Results of a tenyear intervention trial. U.S. Public Health Service Hospitals Cooperative Study Group. Circulation Research 40 (Suppl. I): 98-105, 1977 Medical Research Council Working Party. MRC trial of treatment of mild hypertension: principal results. British Medical Journal 291: 97-104, 1985 Medical Research Council Working Party. MRC trial of treatment of hypertension in older adults: principal results. British Medical Journal 304: 405-412, 1992 Middeke M, Richter WO, Schwand P, Beck B, Holzgreve H. Normalization of lipid metabolism after withdrawal from antihypertensive long-term therapy with beta blockers and diuretics. Arteriosclerosis 10: 145-147, 1990 Middeke M, Weisweiler P, Schwandt R, Holzgreve H. Serum lipoproteins during antihypertensive therapy with beta blockers and diuretics: a controlled long-term comparative trial. Clinical Cardiology 10: 94-98, 1987 Report of the Joint National Committee on detection, evaluation, and treatment of high blood pressure. Archives of Internal Medicine 148: 1023-1038, 1988 Rosenthal J. Antihypertensiver E/Tekt und endokrine Veriinderungen unter Therapie mit Kalziumantagonisten allein sowie in Kombination mit Diuretika. In Distler A, Philipp Th (Eds) Neuere Entwicklungen in der Hochdrucktherapie, pp. 173-180, SchattauerVerlag, Stuttgart, 1983 Salvetti A, Magagna A, Innocenti P, Ponzanelli F, Cagianelli A, et al. The combination of chlorthalidone with nifedipine does not exert an additive antihypertensive e/Tect in essential hypertensives: a crossover multicenter study. Journal of Cardiovascular Pharmacology 17: 332-335, 1991 SHEP Cooperative Research Group. Prevention of stroke by antihypertensive drug treatment in older persons with isolated systolic hypertension. Journal of the American Medical Association 265: 3255-3264, 1991 Veterans Administration Cooperative Study Group on Antihypertensive Agents. E/Tects of treatment on morbidity in hypertension. Results in patients with diastolic blood pressure averaging liS through 129 mm Hg. Journal of the American Medical Association 202: 116121, 1967 Veterans Administration Cooperative Study Group on Antihypertensive Agents. E/Tects of treatment on morbidity in hypertension. II. Results in patients with diastolic blood pressure averaging 90 through 114 mm Hg. Journal of the American Medical Association 213: 11431152, 1970 Veterans Administration Cooperative Study Group on Antihypertensive Agents. E/Tects of treatment on morbidity in hypertension. III. Influence of age, diastolic pressure, and prior cardiovascular disease; further analysis of side e/Tects. Circulation 45: 991-1004, 1972 Wilhelmsen L, Berglund G, Elmfeldt D, Fitzsimons T, Holzgreve H, et al. Beta-blockers versus diuretics in hypertensive men: main results from the HAPPHY trial. Journal of Hypertension 5: 561-572, 1987 Williams WR, Borhani HD, Schnaper HW, Schneider KA, Slatko/TL. The relationship between diuretics and serum cholesterol in HDFP participants. Abstract. Journal of the American College of Cardiology I: 623, 1983 Correspondence and .reprints: Prof. Dr med. Heinrich Holzgreve, University of Munich, Medizinische Poliklinik, Pettenkoferstrasse 8a, 8000 Munich 2, Germany.
Risk-Benefit Aspects of Antihypertensive Drugs H. Holzgreve and M. Middeke University of Munich, Medizinische Poliklinik, Munich, Germany
Summary
Several long term trials using traditional antihypertensive therapy with diuretics and J3-blockers have shown that antihypertensive therapy reduces the overall risk of cardiovascular complications. However, even after several years of therapy the cardiovascular risk in hypertensive patients cannot be lowered to that in the normotensive population. Antihypertensive therapy can reduce the incidence of cerebrovascular complications in patients with hypertension by about 65%. However, the effect of such therapy in preventing coronary events has been disappointing, as these events are 3 to 4 times more common than cerebrovascular complications in hypertensive patients. It is now apparent that adverse pharmacological effects of diuretics and J3-blockers on lipid metabolism persist for many years. Thus, treatment with these agents constitutes a new risk factor for coronary heart disease and may, at least in part, explain the failure of traditional antihypertensive therapy to reduce the incidence of myocardial infarction and sudden death as effectively as that of cerebrovascular accidents. On the other hand, titration of these antihypertensive agents to the lowest possible dose in order to avoid metabolic alterations and subjective adverse effects has frequently resulted in the administration of subtherapeutic doses, particularly for hydrochlorothiazide. Until comparative long term clinical trials with older and newer antihypertensive agents and morbidity and mortality as end-points are completed, the debate on firstline drugs for antihypertensive treatment will not be satisfactorily resolved.
During the past decade, several new drugs have been introduced into antihypertensive therapy, thus providing alternative options for treatment. The multiplicity of substances now available for therapy is naturally a fertile field for debate among those involved in the treatment of hypertension. However, even the recommendations of expert bodies such as the Joint National Committee in the US and the German Hypertension League have undergone a fundamental change during the last 2 decades. During the 1970s and until the mid-1980s, diuretics and /3-blockers were recommended worldwide as first-line drugs for the treatment of hypertension. The above-mentioned organisations currently recommend 4 groups of drugs as being of
equal value: diuretics, /3-blockers, calcium antagonists and angiotensin converting enzyme (ACE) inhibitors (Deutsche Liga zur Bekampfung des hohen Blutdrucks e.V. 1990; Report of the Joint National Committee 1988). One of these classes of agents, the calcium antagonists, was developed in Germany, but for some time these drugs were not commonly accepted there for the treatment of hypertension. This has changed dramatically during recent years (fig. 1). The standard units (millions per year) of /3-blockers and diuretics sold in Germany remained fairly static between 1984 and 1990, but those of ACE inhibitors and calcium antagonists steadily increased. In 1990, the numbers of units of calcium antagonists sold
Drugs 44 (Suppl. 1) 1992
68
2. Lipid Metabolism in Long Term Trials
• j3-Blockers o Diuretics • Calcium antagonists o ACE inhibitors
800
1985
1986
1987
1988
1989
1990
Year Fig. 1. Standard units (in millions) of antihypertensive agents sold per year in Germany between 1984 and 1990. ACE = angiotensin converting enzyme.
were higher than those of ACE inhibitors, diuretics or /3-blockers. Similar trends have been observed in many other countries.
1. Hypertension Intervention Trials There has been much debate in favour of one or another individual drug or substance group_ At times, however, no adequate distinction is made between important, less important, appropriate, inappropriate, or theoretically correct but clinically irrelevant criteria. The selection of antihypertensive agents should be based on evidence that the drug or drugs chosen not only lower blood pressure but also reduce the likelihood of the typical complications of hypertension and thus improve the patient's prognosis. An analysis of the results of the various studies performed to date on the long term efficacy of antihypertensive agents shows that 5 years' antihypertensive drug treatment reduces overall mortality by 33%, cardiovascular mortality by 41%, and the incidence of fatal and nonfatal cerebrovascular events by 51 %. However, the complications of coronary heart disease (i.e. fatal and nonfatal myocardial infarction and sudden cardiac death) are reduced by only 15% and such events occur 3 to 4 times more frequently than cerebrovascular events in patients with hypertension (refer to studies cited in table I).
Overall, data from long term trials (see table I) attest to the outstanding benefit of antihypertensive therapy. However, the results are disappointing as far as coronary heart disease events are concerned, particularly since cardiac complications occur 3 to 4 times as often as cerebrovascular events. Diuretics and /3-blockers were the antihypertensive agents used in these trials and many short term studies (up to several months' duration) have revealed adverse effects of both these drugs on serum lipoproteins (Middeke et al. 1987). The failure of these agents to reduce the incidence of myocardial infarction and sudden death as effectively as the occurrence of cerebrovascular accidents has stimulated interest in lipid metabolism during antihypertensive therapy. Deterioration, normalisation and even improvement in lipoprotein levels have all been observed in a few long term studies of diuretics or /3-blockers over several years (Wilhelmsen et al. 1987; Williams et al. 1983). Possible explanations are that the pharmacological effects on lipoproteins disappear during longer term use, or that other factors influencing serum lipoprotein concentrations, for instance changes in diet, bodyweight or aging, may override any pharmacological effect. Table I. Reduction in mortality and morbidity in major hypertension intervention trials calculated as the percentage reduction in treated groups (extrapolated over 5 years) compared with control or placebo groups [Amery et al. 1985 (EWPHE); Dahlof et al. 1991 (STOP); Hypertension Detection and Follow-up Program Cooperative Group (HDFP) 1979a.b. 1982a.b. 1984a.b. 1985; Management Committee 1980; McFate Smith 1977 (USPH); Medical Research Council Working Party (MRC) 1985. 1992; SHEP Cooperative Research Group 1991; Veterans Administration Cooperative Study Group on Antihypertensive Agents 1967. 1970.1972] Percentage reduction Overall mortality Cardiovascular mortality Cerebrovascular events Coronary heart disease events
33 41 51 15
69
Risk-Benefit Aspects of Antihypertensive Drugs
Patients in the Munich centre who participated in the HAPPHY trial (Wilhelmsen et al. 1987) one of the major randomised long term intervention trials comparing diuretics and {j-blockers provided a unique opportunity to observe the effects of long term treatment and subsequent discontinuation of antihypertensive treatment (Middeke et al. 1990). After 5.2 years of therapy, the mean doses of hydrochlorothiazide and atenolol were 51 and 75mg daily, respectively (table II). Blood pressure was 145/91mm Hg in patients treated with hydrochlorothiazide and 142/ 93mm Hg in patients treated with atenolol; during the 4.6-week post-treatment washout period, blood pressure increased significantly to 165/104 and 159/ 106mm Hg, respectively. After hydrochlorothiazide was discontinued, total cholesterol decreased from 6.4 to 6.0 mmol/L (247 to 231 mg/dl) and low density lipoprotein (LDL) cholesterol from 4.3 to 3.9 mmol/L (167 to 150 mg/dl) [both p = 0.01 vs during treatment], and high density lipoprotein (HDL) cholesterol remained unchanged. Cessation of treatment with the ,a-blocker led to a decrease in LDL cholesterol from 4.2 to 3.9 mmol/L (162 to 150 mg/dl) and an increase in HDL cholesterol from 0.96 to 1.17 mmol/L [37 to 45 mg/dl], both of which were significant changes. These results clearly demonstrate that the lipid changes were reversible after medication was discontinued. On the other hand, the adverse effects of the diuretic and the ,a-blocker had persisted for more than 5 years. Thus, treatment with these agents has introduced a risk factor for coronary heart disease. This increased risk of cardiovascular events may well have reduced the benefit of treatment.
3. Comparison 0/ a Diuretic and a Calcium Antagonist A consequence of these results has been recommendations of dosage reductions, as many side effects of drugs are dose dependent. Whereas 10 to 15 years ago clinicians used daily doses of up to 200mg of hydrochlorothiazide, doses of only 25 or at most 50mg daily are currently recommended. Since the dose-response relationship was claimed
Table II. Blood pressure and serum lipoproteins after long term therapy (5.2 years) with the diuretic hydrochlorothiazide (HCT) or the /3-blocker atenolol. and 4.6 weeks after discontinuation of treatment (Middeke et al. 1990) Diuretic (HCT 51 mg/day)
/3-Blocker (atenolol 75 mg/day)
Number of patients
23
17
Blood pressure (mm Hg) Long term therapy Discontinuation
145/91 165/104*
142/93 159/106*
Total cholesterol (mmol/L) [mg/dl] Long term therapy Discontinuation
6.4 [247] 6.0 [231]**
6.2 [240] 6.0 [233]
LDL cholesterol (mmoI/L) [m9ldl] Long term therapy Discontinuation
4.3 [167] 3.9 [150]**
4.2 [162] 3.9 [150]**
HDL cholesterol (mmol/L) [ma/dl] Long term therapy Discontinuation
1.19 [46] 1.17 [45]
0.96 [37] 1.17 [45]·**
= high density lipoprotein; LDL = low density lipoprotein. * p = 0.001 vs end of therapy; *. p = 0.01 vs end of therapy; ••* p = 0.05 vs end of therapy.
Abbreviations: HDL
to be flat, especially for diuretics, reducing the dosage was not expected to reduce the antihypertensive effect. However, efficacy is dose dependent and it may well be that dose reduction has resulted in administration of less effective doses. For this reason, a study was initiated to evaluate a thiazide diuretic and a calcium antagonist. The I-year comparative therapeutic trial in 369 patients with mild to moderate hypertension (Holzgreve et al. 1989) had a multicentre randomised double-blind design with a single-blind placebo baseline period of 2 weeks. Patients who had been previously treated with antihypertensive agents had an additional 2-week washout period. Hydrochlorothiazide, which has been used in many intervention trials and which is still taken as the gold standard in antihypertensive therapy, was the selected diuretic, and the comparative drug, verapamil, was selected as an example of agents so far
70
Drugs 44 (Suppl. 1) 1992
not used in long term intervention trials in hypertension. Dosage levels of hydrochlorothiazide used were 12.5mg, 25mg and 50mg daily, and verapamil was given at a dosage of 120,240 and 480mg daily. The target blood pressure was defined as a diastolic blood pressure of less than 90mm Hg. If this was not achieved after 2 weeks the next regimen was prescribed. Where even the highest of the 3 dose steps did not result in achievement of target blood pressure, the 2 drugs were combined in 2 dose steps, i.e. initially hydrochlorothiazide 25mg plus verapamil 240mg once daily increasing to 50mg plus 480mg, respectively, on dose step V if required. According to the protocol, patients with mild, moderate and severe hypertension were included. Of the 369 patients who entered the trial, 187 were randomised to treatment with hydrochlorothiazide and 182 to the verapamil group. The randomisation resulted in good comparability between the 2 treatment groups, and both groups had similar systolic and diastolic blood pressures. Mean systolic blood pressures were 152.6 and 152.3mm Hg, in the hydrochlorothiazide and verapamil groups, respectively, and mean diastolic blood pressures were 104.4 and 104.2mm Hg, respectively.
4. Response Rates 4.1 Monotherapy At 8 weeks, the percentage of patients at the target blood pressure, i.e. diastolic less than 90mm Hg, was 42.7% for hydrochlorothiazide-treated patients and 57.7% for verapamil recipients (p < 0.005; fig. 2). At 24 weeks, the values were 26.4% and 46.8% (p < 0.0001) and after 48 weeks 24.7% and 44.8% (p < 0.0001), respectively. These differences show that verapamil significantly was more effective than hydrochlorothiazide at all time points during the study. Furthermore, verapamil was more effective at all dose steps used. Almost twice as many patients on monotherapy with verapamil reached the end of the study compared with monotherapy with hydrochlorothiazide. The fact that more patients remained for a longer time on monotherapy with verapamil is important to consider when analysing the adverse effects of the 2 drugs. In or-
der to determine whether individual patients responded better to one or the other of the medications, subgroup analyses were performed, and comparative results of patients with a blood pressure at entry below and above the median of 103mm Hg, patients aged below 45, 45 to 55 and above 55 years of age, and patients from clinical outpatient departments and from private practices are also shown in figure 2. In all subgroups, patients treated with verapamil had a higher response rate than patients treated with hydrochlorothiazide. Thus, verapamil was more effective than hydrochlorothiazide in all subgroups. 4.2 Combination Therapy It has been suggested that combinations of calcium antagonists and diuretics (or related compounds) are of little value in the treatment of hypertension. Indeed, in several studies there was no significant blood pressure-lowering effect from adding diuretics such as bendroflumethiazide (Cappuccio et al. 1986, 1987, 1991) mefruside (Rosenthal 1983) or chlorthalidone (Salvetti et al. 1991) to nifedipine or amlodipine. In one study only, the addition of hydrochlorothiazide to verapamil caused a further significant fall in blood pressure (MacGregor et al. 1988). Furthermore, it has been postulated that there is a sequential effect, i.e. the antihypertensive effect is improved when a calcium antagonist is added to a diuretic, but not when a diuretic is added to a calcium antagonist (Hansson 1991). As already shown in our study, 24.7% of patients achieved target blood pressure after 48 weeks' monotherapy with hydrochlorothiazide, whereas the success rate after verapamil monotherapy was 44.8% (fig. 3), i.e., when the drugs were used as monotherapy, verapamil was effective almost twice as often as hydrochlorothiazide. Patients who failed to respond to monotherapy were given verapamil and hydrochlorothiazide in combination. The addition of verapamil to patients who were initially randomised to hydrochlorothiazide and did not achieve target blood pressure increased the response in 34.3% of patients. Conversely, addition
Risk-Benefit Aspects of Antihypertensive Drugs
71
70 60 iii'
E 50
.~
iii
"-
'0 40
t
Q> V>
c
30
0
"V> 20 Q>
a:
10 0
All
patients
103 > 103 BP (mm Hg)
~
45-55 > 55 Years 01 age
< 45
In In private clinics practice
Fig. 2. Percentages of patients (all patients and various subgroups) achieving target blood pressure after 8 weeks' monotherapy with hydrochlorothiazide (0) or verapamil (.).
of hydrochlorothiazide to patients who were initially randomised to verapamil and did not achieve target blood pressure increased the response rate in only 17.8% of patients; thus, compared with hydrochlorothiazide, verapamil was not only effective twice as often when used as monotherapy but also when used as a drug for combination. These results clearly show that there is an additive effect from combining verapamil and hydrochlorothiazide. Thus, according to the previous studies there might be a difference between calcium antagonists SUQb as verapamil and the dihydropyridines. In cSu~ study, the antihypertensive efficacy of hydrochlorothiazide was lower than that of verapamil, and more patients achieved target blood pressure after the addition of verapamil to hydrochlorothiazide compared with adding the diretic to the calcium antagonist. It is possible that indings from previous studies that failed to dem.nstrate a statistically significant effect of diuretics n patients already taking a calcium antagonist night have been influenced by the small populaion studied (Cappuccio et al. 1986, 1987, 1991), lr may have resulted from the use of subtherapeuic diuretic doses (Salvetti et al. 1991). As far as a possible sequential effect is concerned, it was to be expected that the total percentage of patients who responded either to monotherapy or combination
therapy would be the same in both groups, assuming an additive effect and excluding any sequential effect. Indeed, the figure for total response rate was almost identical, i.e. 59.0% in the hydrochlorothiazide group and 62.6% in the verapamil group (fig. 3). This result excludes any sequential effect.
5. Adverse Effects and Withdrawals The higher efficacy of verapamil might be due to the fact that in recent years titration to the lowest possible dose of diuretics has been suggested in order to avoid side effects and adverse metabolic alterations, such as those on blood lipoprotein leviii' 75 C
59.0%
0; "- 50
Plus
62.6%
.~
'0 ~
VER
17.8% 34.3% Mono
Q> V>
c 25 0 a.
VER
44.8%
V> Q>
a:
0
HCT group
Verapamil group
Fig. 3. Percentages of patients achieving target blood pressure in groups randomised to 48 weeks' therapy with hydrochlorothiazide (HCT) or verapamil (VER) as monotherapy or the 2 drugs combined.
Drugs 44 (Suppl. 1) 1992
72
els. In the current study, this strategy may well have resulted in administration of doses that cause fewer adverse effects but that reduce the elevated blood pressure less effectively. On the other hand, the doses ofverapamil selected in our study might well be more potent in reducing blood pressure, but at the same time could cause more adverse effects than hydrochlorothiazide. The answers provided by the patients in this double-blind study using a structured questionnaire show that the patients treated with hydrochlorothiazide complained relatively more commonly of urinary frequency and those treated with verapamil reported constipation more frequently, although the differences were not analysed for statistical significance. Potassium salts and potassium-sparing diuretics were prescribed more often for patients in the hydrochlorothiazide group than for patients in the verapamil group. The only significant change in laboratory parameters was a mean fall of 0.311 mmol/L in serum potassium. However, contrary to findings from our earlier studies, in the group of 178 patients receiving hydrochlorothiazide no significant changes in total cholesterol, LDL cholesterol, HDL cholesterol and triglycerides could be observed at 8 weeks. It is suggested that this was due to the lower mean daily dose of hydrochlorothiazide used (32mg) compared with a dose of 51 mg used in our studies cited above (Middeke et al. 1990). The total number of patients withdrawing from the study are shown in figure 4. Administrative reasons such as noncompliance or protocol violations accounted for 30 to 65% of withdrawals across the 4 groups. Increased diastolic blood pressure (> 120mm Hg) was the most frequent adverse effect causing withdrawal among all remaining patients. Similar number of patients were withdrawn from each monotherapy group. However, it must be remembered that because of the greater efficacy of verapamil, more patients remained on verapamil for a longer time than on hydrochlorothiazide and therefore were at a higher risk of adverse effects associated with verapamil. There were 10 withdrawals by patients for whom verapamil had been added to hydrochlorothiazide and 17 dropouts in
"'c:.
25
~ a.
20
ci
15
"5
.s $
~
iii ~
~
25
23
17
10 10
5
"0
-s ~
0
HCT
VER
Monotherapy
HCT
+ VER
VER + HCr
Combination therapy
Fig. 4. Total number of patients withdrawing from randomised therapy during the trial, on monotherapy with hydrochlorothiazide or verapamil or after adding verapamil to hydrochlorothiazide (HCT + VER) or hydrochlorothiazide to verapamil (VER + HCT).
patients who had hydrochlorothiazide added to verapamil therapy. Thus, the withdrawal rate during treatment with verapamil seems no higher than that during treatment with hydrochlorothiazide.
6. Conclusions and Therapeutic Implications In conclusion, findings from this study have shown that smaller doses of the diuretic reduce the incidence of side effects, particularly of adverse lipid alterations, and that the incidence of treatmentrelated adverse effects and withdrawals was essentially the same with the 2 drugs. However, verapamil had a significantly more pronounced antihypertensive effect than hydrochlorothiazide. Physicians should base their therapeutic decisions on currently available data, although even experts may differ in their interpretation. As mentioned above, expert bodies such .as the German Hypertension League and the Joint National Committee currently recommend 4 groups of drugs as first-line agents: diuretics, {j-blockers, calcium antagonists and ACE inhibitors. However, the results presented here must call these recommendations into question - at least as far as diuretics and calcium antagonists are concerned - since they showed that with regard to antihypertensive efficacy and tolerability standard doses of the calcium antagonist were clearly superior to those of the diuretic.
Risk-Benefit Aspects of Antihypertensive Drugs
References Amery A, Birkenhager W, Brixko p, Bulpitt C, Clement D, et al. Mortality and morbidity results from the European Working Party on High Blood Pressure in the Elderly trial. Lancet I: 1349-1354, 1985 Cappuccio FP, Markandu ND, Singer DRJ, Buckley MG, Miller MA, et al. A double-blind crossover study of the e/Tect of concomitant diuretic therapy in hypertensive patients treated with amlodipine. American Journal of Hypertension 4: 297-302, 1991 Cappuccio FP, Markandu ND, Tucker F, Sagnella GA, MacGregor GA. Does a diuretic cause a further fall in blood pressure in hypertensive patients already on nifedipine? Journal of Clinical Hypertension 4: 346-353, 1986 Cappuccio FP, Markandu ND, Tucker FA, Shore AC, MacGregor GA. A double-blind study of the blood pressure-lowering e/Tect of a thiazide diuretic in hypertensive patients already on nifedipine and a beta-blocker. Journal of Hypertension 5: 733-738, 1987 Dahlbf B, Lindholm LH, Hansson L, Scherstein B, Ekbom T, et al. Morbidity and mortality in the Swedish trial in old patients with hypertension (STOP - Hypertension). Lancet 338: 1281-1285,1991 Deutsche Liga zur Bekiimpfung des hohen Blutdrucks e.V. Empfehlungen zur Hochdruckbehandlung in der Praxis und zur Behandlung hypertensiver Notfalle, 9th Ed., 1990 Hansson L. Position of calcium antagonists in antihypertensive combination therapy. Abstract. 2nd International Symposium on Calcium Antagonists in Cardiovascular Care, Basle, 1991 Holzgreve H, Distler A, Michaelis J, Philipp T, Wellek S. Verapamil versus hydrochlorothiazide in the treatment of hypertension: results of a long term double blind comparative trial. British Medical Journal 299: 881-886, 1989 Hypertension Detection and Follow-up Program Cooperative Group. Five-year findings of the Hypertension Detection and Follow-up Program. I. Reduction in mortality of persons with high blood pressure, including mild hypertension. Journal of the American Medical Association 242: 2562-2571, 1979a Hypertension Detection and Follow-up Program Cooperative Group. Five-year findings of the Hypertension Detection and Follow-up Program. II. Mortality by race, sex and age. Journal of the American Medical Association 242: 2572-2577, 1979b Hypertension Detection and Follow-up Program Cooperative Group. Five-year findings of the Hypertension Detection and Follow-up program. III. Reduction in stroke incidence among persons with high blood pressure. Journal of the American Medical Association 247: 633-638, 1982a Hypertension Detection and Follow-up Program Cooperative Group. The e/Tect of treatment on mortality in 'mild' hypertension. New England Journal of Medicine 307: 976-980, 1982b Hypertension Detection and Follow-up Program Cooperative Research Group. The e/Tect of antihypertensive drug treatment on mortality in the presence of resting electrocardiographic abnormalities at baseline: the HDFP experience. Circulation 79: 996-1003, 1984a Hypertension Detection and Follow-up Program Cooperative Group. E/Tect of stepped care treatment on the incidence of myocardial infarction and angina pectoris. 5-year findings. Hypertension 6 (Suppl. I): 198-206, 1984b Hypertension Detection and Follow-up Program Cooperative Group. Five-year findings of the Hypertension Detection and Follow-up Program. Prevention and reversal of left ventricular hypertrophy with antihypertensive drug therapy. Hypertension 7: 105-112, 1985 MacGregor GA, Markandu ND, Singer DRJ. Diuretics and calcium antagonists in hypertension. British Journal of Clinical Practice 60: 58-59, 1988
73
Management Committee. The Australian Therapeutic Trial in Mild Hypertension. Lancet I: 1261-1267, 1980 McFate-Smith W. Treatment of mild hypertension. Results of a tenyear intervention trial. U.S. Public Health Service Hospitals Cooperative Study Group. Circulation Research 40 (Suppl. I): 98-105, 1977 Medical Research Council Working Party. MRC trial of treatment of mild hypertension: principal results. British Medical Journal 291: 97-104, 1985 Medical Research Council Working Party. MRC trial of treatment of hypertension in older adults: principal results. British Medical Journal 304: 405-412, 1992 Middeke M, Richter WO, Schwand P, Beck B, Holzgreve H. Normalization of lipid metabolism after withdrawal from antihypertensive long-term therapy with beta blockers and diuretics. Arteriosclerosis 10: 145-147, 1990 Middeke M, Weisweiler P, Schwandt R, Holzgreve H. Serum lipoproteins during antihypertensive therapy with beta blockers and diuretics: a controlled long-term comparative trial. Clinical Cardiology 10: 94-98, 1987 Report of the Joint National Committee on detection, evaluation, and treatment of high blood pressure. Archives of Internal Medicine 148: 1023-1038, 1988 Rosenthal J. Antihypertensiver E/Tekt und endokrine Veriinderungen unter Therapie mit Kalziumantagonisten allein sowie in Kombination mit Diuretika. In Distler A, Philipp Th (Eds) Neuere Entwicklungen in der Hochdrucktherapie, pp. 173-180, SchattauerVerlag, Stuttgart, 1983 Salvetti A, Magagna A, Innocenti P, Ponzanelli F, Cagianelli A, et al. The combination of chlorthalidone with nifedipine does not exert an additive antihypertensive e/Tect in essential hypertensives: a crossover multicenter study. Journal of Cardiovascular Pharmacology 17: 332-335, 1991 SHEP Cooperative Research Group. Prevention of stroke by antihypertensive drug treatment in older persons with isolated systolic hypertension. Journal of the American Medical Association 265: 3255-3264, 1991 Veterans Administration Cooperative Study Group on Antihypertensive Agents. E/Tects of treatment on morbidity in hypertension. Results in patients with diastolic blood pressure averaging liS through 129 mm Hg. Journal of the American Medical Association 202: 116121, 1967 Veterans Administration Cooperative Study Group on Antihypertensive Agents. E/Tects of treatment on morbidity in hypertension. II. Results in patients with diastolic blood pressure averaging 90 through 114 mm Hg. Journal of the American Medical Association 213: 11431152, 1970 Veterans Administration Cooperative Study Group on Antihypertensive Agents. E/Tects of treatment on morbidity in hypertension. III. Influence of age, diastolic pressure, and prior cardiovascular disease; further analysis of side e/Tects. Circulation 45: 991-1004, 1972 Wilhelmsen L, Berglund G, Elmfeldt D, Fitzsimons T, Holzgreve H, et al. Beta-blockers versus diuretics in hypertensive men: main results from the HAPPHY trial. Journal of Hypertension 5: 561-572, 1987 Williams WR, Borhani HD, Schnaper HW, Schneider KA, Slatko/TL. The relationship between diuretics and serum cholesterol in HDFP participants. Abstract. Journal of the American College of Cardiology I: 623, 1983 Correspondence and .reprints: Prof. Dr med. Heinrich Holzgreve, University of Munich, Medizinische Poliklinik, Pettenkoferstrasse 8a, 8000 Munich 2, Germany.
